Professional Documents
Culture Documents
Linear System Theory and Design - Chi-Tsong Chen PDF
Linear System Theory and Design - Chi-Tsong Chen PDF
Theorem 5.1 A discrete-time SISO system described by (5.7) is BIBO stable if and only if gk] is absolutely summable in (0, 20) or Distt0. approaches &(1) +a 2. The output excited by u[k] = sin wok, for k > 0, approaches [g(e/)| sin(wok+ K8(e/*)) where is the z-transform of g[k] or BO =D abe 68) = Proof: If u[k] for all k > O, then (5.7) becomes ‘ ‘ yl] = DO glue =m) = a gtr] o ra which implies yik] > a Y* gi] = ag(l) ask + 00 om where we have used (5.8) with z= 1. This establishes the first part of Theorem 5.D2. If u{k] = sin cok, then (5.7) becomes k yl = Yo gl] sin walk — m) mo k = SY sim [sin wk cos wm — cos.aigk sina) noo * ‘ = sin wok S° gl] coswom — cos wok J gm) sin wom nod mao Thus we have, as k + 00. Lk] + sin wok J gm] cos.wom ~ cos wk 5 > gin] sin arom (59) no no If g{k] is absolutely summable, we can replace = by e/* in (5.8) to yield DY simle" = J glmlicos om — j sinwm] = m= &e!*)128 STABILITY Thus (5.9) becomes YE} > sin ook (Re[g(e/")]) + cos wok (Im[g(e/")}) = [Ble!)i sinwok+ 8 le/)) This completes the proof of Theorem 5.D2. Q.E.D. Theorem 5.D2 is a basic result in digital signal processing. Next we state the BIBO stability in terms of discrete proper rational transfer functions. Theorem 5.03, A discrete-time SISO system with proper rational transfer function (2) is RIRO stable if and only if every pole of 85) has a magnitude less than | of equivalently, lies inside the unit circle on the <-plane 1f 8(2) has pole p, with multiplicity m,, then its partial fraction expansion contains the factors =p = pom ‘Thus the inverse z-transform of &(z) or the impulse response sequence contains the factors BG Phe kph oe, [Ris straightforward to verify that every such term is absolutely summable if and only if p, has 4 magnitude less than I. Using this fact, we can establish Theorem 5.3, Inthe continuous-time case, an absolutely integrable function f(r), as shown in Fig. 5.1, may not be bounded and may not approach zero as 1 —» oc. In the discrete-time case, if g{K] is absolutely summable, then it must be bounded and approach zero as k + 20, However, the converse is not tre as the next example shows. ExaMPLe 5.3 Consider a discrete-time LTI system with impulse response sequence g{k] = I/k for k = 1, 2,..., and g{0] = 0. We compute Leek =; = & el ee ‘)..(! 1 i se a) oe i) ‘There are two terms, each is } or larger, in the first pair of parentheses; therefore their sum is larger than }, There are four terms, each is } or larger, in the second pair of parentheses: therefore their sum is larger than $. Proceeding forward we conclude s tot 20 This impulse response sequence is bounded and approaches (as k —+ 90 but is not absolutely summable. Thus the discrete-time system is not BIBO stable according to Theorem 5.D1. The transfer function of the system can be shown to equal 5.3 Internal Stability 129 A= +74 tis not a rational function of z and Theorem 5.D3 is not applicable. For multivariable discrete-time systems, we have the following results, Theorem 5.MD1 A MIMO discrete-time system with impulse response sequence matrix G[k] = [gi){k]] is BIBO stable if and only if every gi, [A] is absolutely summable. Theorem 5.MD3 [A MIMO discrete-time system with discrete proper rational transfer matsix G(:) = [8 (2)] is BIBO Stabe if and only if every pole of every (z) his a magnitude les than 1 We now discuss the BIBO stability of discrete-time state equations. Consider x(k + 1] = ANlA] + Bul] (5.10) ylk] = Cx(k) + Dulk] Its discrete transfer matrix is Gq) = C= Ay'B+D ‘Thus Equation (5.10) or, to be more precise, the zero-state response of (5.10) is BIBO stable if and only if every pole of G(z) has a magnitude less than 1. We discuss the relationship between the poles of G(z) and the eigenvalues of A. Because of 1 ———~—__ C[Adj(z1 — A)]B+D Sp ClAdict~ a every pole of G(=) is an eigenvalue of A. Thus if every eigenvalue of A has a negative real part, then (5.10) is BIBO stable. On the other hand, even if A has some eigenvalues with zero or positive real part. (5,10) may, as in the continuous-time case, still be BIBO stable. 5.3 Internal Stability ‘The BIBO stability is defined for the zero-state response. Now we study the stability of the zero-input response or the response of Axi) (sly x) excited by nonzero initial state xp. Clearly, the solution of (5.11) is 6.12)130 STABILITY Definition 5.1 The zero-input response of (5.4) or the equation X = Ax is marginally stable or stable in the sense of Lyapunov if every finite initial state x, excites a bounded response. Itisasymptotically stable if every finite initial state excites a bounded response, which, in addition, approaches 0.as t —> 00. ‘We mention that this definition is applicable only to linear systems. The definition that is applicable to both linear and nonlinear systems must be defined using the concept of equivalence states and can be found, for example, in Reference (6, pp. 401-403]. This text studies only linear systems: therefore we use the simplified Definition 5.1 Theorem 5.4 1 1. The equation X = Ax is marginally stable if and only if all eigenvalues of A have zero or negative zeal parts and those with zero real parts are simple roots of the minimal polynomial of A. 2. The equation & parts AX is asymptotically stable if and only if all eigenvatues of A have negative real We first mention that any (algebraic) equivalence transformation will not alter the stability of a state equation. Consider = Px, where P is a nonsingular matrix. Then X = AX is equivalent to X = AX = PAP~'X. Because P is nonsingular, if x is bounded, so is &; if x approaches 0 as ¢ + 00, so does X. Thus we may study the stability of A by using A. Note that the eigenvalues of A and of A are the same as discussed in Section 4.3. The response of & = AX excited by €(0) equals X(t) = e4x(0). It is clear that the rTesponse is bounded if and only if every entry of e*’ is bounded for all ¢ > 0. If A is in Jordan form, then e¥ is of the form shown in (3.48). Using (3.48), we can show that if an eigenvalue has a negative real part, then every entry of (3.48) is bounded and approaches 0 as 1 —+ oo, Jf an eigenvalue has zero real part and has no Jordan block of order 2 or higher, then the corresponding entry in (3.48) is a constant or is sinusoidal for all t and is, therefore, bounded. ‘This establishes the sufficiency of the first part of Theorem 5.4. If A has an eigenvalue with a Positive real part, then every entry in (3.48) will grow without bound. If A has an eigenvalue with zero real part and its Jordan block has order 2 or higher, then (3.48) has at least one entry that grows unbounded. This completes the proof of the first part. To be asymptotically stable, every entry of (3.48) must approach zero as t —> 00. Thus no eigenvalue with zero real partis, Permitted. This establishes the second part of the theroem, Examte 5.4 Consider 00 0 x=]00 0 |x 00-1 Its characteristic polynomial is A(4) = 47(. + 1) and its mimimal polynomial is yo) = AG. + 1). The matrix has eigenvalues 0, 0, and —I. The eigenvlaue 0 is a simple root of the minimal polynomial. Thus the equation is marginally stable. The equation 5.3 Internal Stability 131 01 0 00 0 |x 00 is not marginally stable, however, because its minimal polynomial is 4°(A + 1) and 4 = Ois not a simple root of the minimal polynomial. As discussed earlier, every pole of the transfer matrix GG) = CI-A)y'B+D is an eigenvalue of A. Thus asymptotic stability implies BIBO stability. Note that asymptotic stability is defined for the zero-input response, whereas BIBO stability is defined for the zero-state response. The system in Example 5.2 has eigenvalue 1 and is not asymptoti- cally stable; however, it is BIBO stable. Thus BIBO stability, in general, does not im- ply asymptotic stability. We mention that marginal stability is useful only in the design of oscillators. Other than oscillators, every physical system is designed to be asymptoti- cally stable or BIBO stable with some additional conditions, as we will discuss in Chap- ter7. 5.3.1 Discrete-Time Case This subsection studies the internal stability of discrete-time systems or the stability of xfk +1) = Axtk] (6.13) excited by nonzero initial state xo. The solution of (5.13) is, as derived in (4.20), xlk] = Atx, 6.14) Equation (5.13) is said to be marginally stable or stable in the sense of Lyapunov if every finite initial state x, excites a bounded response. It is asymptotically stable if every finite initial state excites a bounded response. which, in addition, approaches 0 as k > 00. These definitions are identical to the continuous-time case Theorem 5.D4 1. The equation x{k + 1] = Ax[k] is marginally stable if and only if all eigenvalues of A have ‘magnitudes less than or equal to | and those equal to 1 are simple roots of the minimal polynomial of A 2. The equation x{k + 1] = Ax{K] is asymptotically stable if and only if all eigenvalues of A have magnitudes less than 1 ‘As in the continuous-time case, any (algebraic) equivalence transformation will not alter the stability of a state equation. Thus we can use Jordan form to establish the theorem. The proof is similar to the continuous-time case and will not be repeated. Asymptotic stabiSTABILITY implies BIBO stability but not the converse. We mention that marginal stability is useful only in the design of discrete-time oscillators. Other than oscillators. every discrete-time physical system is designed to be asymptotically stable or BIBO stable with some additional conditions, as we will discuss in Chapter 7. 5.4 Lyapunov Theorem This section introduces a different method of checking asymptotic stability of X = Ax. For convenience, we call A stable if every eigenvalue of A has a negative real part. Theorem 5.5 All eigenvalues of A have negative real pans'it and only if for any given positive definite symmetric matrix N, the Lyapunov equation A‘M+MA=-N (5.15) hhas a unique symmetric solution M and M is positive definite, Corollary 5.5 All eigenvalues of an mx m matrix A have ni N with m < 1 and with the property -gative real parts if and only if for any given m % nt matrix rank O := rank = allcolamn rank) 5.6) Nan were Os anim x m mati. the Lyapunov equation AM+MA=-NN 6.17) has a unique symmetric solution M and M is positive defini. For any N, the matrix N in (5.17) is positive semidefinite (Theorem 3.7). Theorem 5.5 and its corollary are valid for any given N: therefore we shall use the simplest possible N. Even so. using them to check stability of A is not simple. It is much simpler to compute. using MATLAB, the eigenvalues of A and then check their real parts. Thus the importance of Theorem 5.5 and its corollary is not in checking the stability of A but rather in studying the stability of nonlinear systems. They are essential in using the so-called second method of Lyapunov. We mention that Corollary 5.5 can be used to prove the Routh-Hurwit test. See Reterence [6. pp. 417-419] Proof of Theorem 5.5. Necessity: Equation (5.15) is a special case of (3.59) with A = A” and B = A. Because A and A’ have the same set of eigenvalues. if A is stable, A has no two eigenvalues such that A, +4) = 0. Thus the Lyapunov equation is nonsingular and. has @ unique solution M for any N. We claim that the sotution can be expressed as 5.4 Lyapunov Theorem 133 y f eM'NeMat (5.18) Indeed, substituting (5.18) into (5.15) yields AM+MA fo aeiner ar f eNeMA dr 0 i =0- (5.19) where we have used the fact ef = 0 att = 2 for stable A. This shows that the M in (5.18) is the solution. It is clear that if N is symmetric, so is M. Let us decompase N as N = NN, where N is nonsingular (Theorem 3.7) and consider x’Mx i xe'N'Ne“xdt f Ne“ x) 3 (5.20) b jar, for any nonzero x. the int Because both N and e¥ are nonsing ‘grand of (5.20) is positive for every r. Thus x'Mx is positive for any x 4 0. This shows the positive definiteness of M. Sufficiency: We show that if N and M are positive definite, then A is stable. Let 4 be an eigenvalue of A and v # Obe a corresponding eigenvector: that is. AV = Av. Even though Aisareal matris, its eigenvalue and eigenvector can be complex. as shown in Example 3.6. ‘Taking the complex-conjugate transpose of Av = AV yields v"A* = v"A\ where the asterisk denotes complex-conjugate transpose. Premultiplying v* and postmultiplying ¥ to (5.15) yields =v'Nv = v‘A/My + v"MAv = ("+ ivMy = 2Re(ary"My Because v'Mv and y°Nv are, as discussed in Section 3.9, both real and positive, implies Re() < 0. This shows that every eigenvalue of A has a negative real part QED. The proof of Corollary 5.5 follows the proof of Theorem 5.5 with some modification. We discuss only where the proof of Theorem 5.5 is not applicable. Consider (5.20). Now N ism x with m You might also like
An Introduction To Mixed-Signal IC Test and Measurement (Burns)Document352 pagesAn Introduction To Mixed-Signal IC Test and Measurement (Burns)Gatocontos OMNo ratings yet Consulting Math DrillsDocument15 pagesConsulting Math DrillsHashaam Javed50% (2) Linear System TheoryDocument5 pagesLinear System Theoryfrmshibu100% (1) White and Woodson PDFDocument330 pagesWhite and Woodson PDFAnonymous ivCODLhn5No ratings yet An Introduction To Signal Detection and Estimation 1988Document410 pagesAn Introduction To Signal Detection and Estimation 1988Rasool Faraji75% (4) All Art Is Propaganda PDFDocument2,708 pagesAll Art Is Propaganda PDFPau_Brand_255033% (3) (Wai-Kai Chen) Passive, Active, and Digital Filter (Bookos - Org) BiquadDocument26 pages(Wai-Kai Chen) Passive, Active, and Digital Filter (Bookos - Org) BiquadThulasi ReddyNo ratings yet Fundamentals of Lyapunov TheoryDocument9 pagesFundamentals of Lyapunov Theorygoitom01No ratings yet Reference Frame TheoryDocument5 pagesReference Frame Theoryaravi1979No ratings yet Van LoanDocument273 pagesVan LoanLeticia TonettoNo ratings yet Dynamic Model of Induction Motors For Vector ControlDocument10 pagesDynamic Model of Induction Motors For Vector ControlMarica RalucaNo ratings yet Matrix ExponentialDocument14 pagesMatrix Exponentialvignesh0617No ratings yet Thyristor DC Drives. P.c.senDocument307 pagesThyristor DC Drives. P.c.senpalu259478% (9) Feedback Control Systems ChaptDocument23 pagesFeedback Control Systems ChaptJavier ToscanoNo ratings yet Mod12 - Lecture 1Document31 pagesMod12 - Lecture 1VAISHAKA N RAJ100% (1) Symmetrical ComponentsDocument8 pagesSymmetrical Componentssrikanth velpulaNo ratings yet BOOK4 PsDocument128 pagesBOOK4 Psgoyo2kNo ratings yet LQR FeedforwardDocument20 pagesLQR Feedforwardsri248No ratings yet r05410205 Advanced Control SystemsDocument8 pagesr05410205 Advanced Control SystemsShashidhar KasthalaNo ratings yet 1980 - Stable Adaptive Controller Design, Part II - Proof of Stability (Narendra)Document9 pages1980 - Stable Adaptive Controller Design, Part II - Proof of Stability (Narendra)ademargcjuniorNo ratings yet bjorckLSbSolving Linear Least Squares Problems by Orthogonal Gramschmidt OrthogonalizationDocument21 pagesbjorckLSbSolving Linear Least Squares Problems by Orthogonal Gramschmidt Orthogonalizationdgh3No ratings yet Solution To The Positive Real Control Problem For Linear Time-Invariant SystemsDocument13 pagesSolution To The Positive Real Control Problem For Linear Time-Invariant SystemsRohit GandhiNo ratings yet Complex Vector Model of The Squirrel Cage Induction Machine Including Instantaneous Rotor Bar CurrentsDocument8 pagesComplex Vector Model of The Squirrel Cage Induction Machine Including Instantaneous Rotor Bar CurrentsJorge Luis SotoNo ratings yet MatlabDocument45 pagesMatlabHamid SiNo ratings yet Single Sideband ModulationDocument25 pagesSingle Sideband ModulationADITYA GAURNo ratings yet Overview of Power System StabilityDocument21 pagesOverview of Power System StabilitySanjayan SubramaniamNo ratings yet Emanuel Parzen Modern Probability Theory and Its ApplicationsDocument480 pagesEmanuel Parzen Modern Probability Theory and Its ApplicationsAlfonzo Orozco100% (1) Yury V. Orlov - Discontinuous Systems - Lyapunov Analysis and Robust Synthesis Under Uncertainty Conditions-Springer-Verlag London (2009) PDFDocument333 pagesYury V. Orlov - Discontinuous Systems - Lyapunov Analysis and Robust Synthesis Under Uncertainty Conditions-Springer-Verlag London (2009) PDFJosé Ramón MezaNo ratings yet 09 - Chapter 1Document20 pages09 - Chapter 1Dr. POONAM KAUSHALNo ratings yet Excercise in FDocument2 pagesExcercise in FPaz Villahoz100% (2) ElvisDocument1 pageElvismaui3No ratings yet Calculus Cheat Sheet DerivativesDocument4 pagesCalculus Cheat Sheet DerivativesRajatNo ratings yet Ibfc 142 Engenheiro MecanicoDocument16 pagesIbfc 142 Engenheiro MecanicoJosé Maria junior100% (2) Enfermeiro Ebsrh 2015Document10 pagesEnfermeiro Ebsrh 2015Neila Reis da SilvaNo ratings yet Tema 6. CULTURADocument7 pagesTema 6. CULTURAMarinaNo ratings yet 5756867Document1,027 pages5756867Sách Việt Nam50% (2) Aural RitmoDocument1 pageAural RitmofabioNo ratings yet Planes APIDocument10 pagesPlanes APIManuel FigueroaNo ratings yet Music To Watch Girls by MJDocument3 pagesMusic To Watch Girls by MJMarvin JongNo ratings yet 5Document2 pages5Abel Salazar PianoNo ratings yet Ampacidad AlimentacionDocument1 pageAmpacidad Alimentacionluis miguel sanchez estrellaNo ratings yet Building ActDocument136 pagesBuilding ActVeronika RaušováNo ratings yet Lagenda Violin PDFDocument2 pagesLagenda Violin PDFMarvin Jong0% (1) Flow Design Formulas For Calculation - Exported From (HTM - 02 - 01 - Part - A)Document8 pagesFlow Design Formulas For Calculation - Exported From (HTM - 02 - 01 - Part - A)wesam allabadi50% (2) Tu Me Gustas Trompeta 1Document2 pagesTu Me Gustas Trompeta 1JonathanNo ratings yet Business Startup Guides-Preparing Yourself Ver 2Document4 pagesBusiness Startup Guides-Preparing Yourself Ver 2Minh Huỳnh100% (1) Renault Inyección Electrónica Kangoo RL-RN 1.6 1999 Siemens Sirius 32b PDFDocument2 pagesRenault Inyección Electrónica Kangoo RL-RN 1.6 1999 Siemens Sirius 32b PDFOsvaldo LópezNo ratings yet Hecho Del PíncipeDocument11 pagesHecho Del PíncipeEdisson MoralesNo ratings yet Detail (1-1) To R.C. Strip Footing (Sf1) : A B E F H J KDocument1 pageDetail (1-1) To R.C. Strip Footing (Sf1) : A B E F H J KThomas Garcia0% (1) GM Relés y Fusibles Corsa PDFDocument1 pageGM Relés y Fusibles Corsa PDFcorylinNo ratings yet 01 - Lista de Parametros SimplificadaDocument8 pages01 - Lista de Parametros SimplificadaLuis Felipe VidigalNo ratings yet El Pastor - VihuelaDocument1 pageEl Pastor - VihuelaDamian Guardia SalazarNo ratings yet Greenwashing Nas EmpresasDocument37 pagesGreenwashing Nas EmpresasLara GagoNo ratings yet Escort 18 16 ZetecDocument2 pagesEscort 18 16 ZetecOsvaldo LópezNo ratings yet 388 1006 1 PBDocument20 pages388 1006 1 PBGabriel Evangelista dos SantosNo ratings yet Mix Trompeta ChilombianoDocument1 pageMix Trompeta ChilombianoDenise AlvaradoNo ratings yet 6periostitisDocument10 pages6periostitisAntonioNo ratings yet PDF Parts Catalog Tvs Rockz - CompressDocument104 pagesPDF Parts Catalog Tvs Rockz - CompressaspareteNo ratings yet Renault Inyección Electrónica Clio 1.4 Ac Delco Monopoint P PDFDocument2 pagesRenault Inyección Electrónica Clio 1.4 Ac Delco Monopoint P PDFGood CarNo ratings yet 101 Problems in AlgebraDocument160 pages101 Problems in AlgebraTrishaii Cuaresma96% (28) Algebra 2 Ans PDFDocument2 pagesAlgebra 2 Ans PDFPuja AgarwalNo ratings yet Scripting TCL/PerlDocument5 pagesScripting TCL/PerlRavi Tej GunisettyNo ratings yet Math 9 ReviseDocument13 pagesMath 9 ReviseGem AbetriaNo ratings yet Asn 1Document25 pagesAsn 1Yukti SharmaNo ratings yet Feedback Controller: Proportional, Integral, Derivative (PID)Document34 pagesFeedback Controller: Proportional, Integral, Derivative (PID)Ereen100% (1) Cephalometric Assessment of Sagittal Dysplasia: A Review of Twenty-One MethodsDocument9 pagesCephalometric Assessment of Sagittal Dysplasia: A Review of Twenty-One MethodsKanish AggarwalNo ratings yet واحدها و اندازه گیریDocument31 pagesواحدها و اندازه گیریapi-3706290No ratings yet Autopipe Nozzle Report: Model Name - Sss - NozDocument8 pagesAutopipe Nozzle Report: Model Name - Sss - NozRrrrrrrNo ratings yet Lab Manual of C++Document14 pagesLab Manual of C++Ammar AbbasiNo ratings yet Ap06 FRQ Calculus BC 51710Document6 pagesAp06 FRQ Calculus BC 51710DEVANSHNo ratings yet Physics 71 1st Long ExamDocument12 pagesPhysics 71 1st Long ExamAb YZNo ratings yet Math Grade 3 Module 7/7Document519 pagesMath Grade 3 Module 7/7Paula SippyNo ratings yet CS 301 - Short NotesDocument13 pagesCS 301 - Short NotesMubshera SaddiqaNo ratings yet Secondary Calibration of Acoustic Emission Sensors: Standard Practice ForDocument7 pagesSecondary Calibration of Acoustic Emission Sensors: Standard Practice ForEric GozzerNo ratings yet AssignmentDocument6 pagesAssignmentSwagata SharmaNo ratings yet How To Write An Empirical PaperDocument20 pagesHow To Write An Empirical PaperAra AusejoNo ratings yet Assignment For Prob. and Stat.Document5 pagesAssignment For Prob. and Stat.Jessrell MalacoraNo ratings yet Jurnal QrisDocument14 pagesJurnal Qrishandy uppaNo ratings yet J. Doyne Farmer - Maarten P. Scholl - Anisoara Callinescu - How Market Ecology Explains Market MalfunctionDocument9 pagesJ. Doyne Farmer - Maarten P. Scholl - Anisoara Callinescu - How Market Ecology Explains Market MalfunctionRobertoNo ratings yet JJ207 Thermodynamic Topic 2 First Law of ThermodynamicsDocument34 pagesJJ207 Thermodynamic Topic 2 First Law of ThermodynamicsAh Tiang50% (2) 13 - 1351 Development of Fragility and Reliability CurvesDocument11 pages13 - 1351 Development of Fragility and Reliability CurveshamadaniNo ratings yet AHP SyllabusDocument1 pageAHP SyllabusAjmal SheriffNo ratings yet Modeling Complex Ecological Systems - An IntroductionDocument6 pagesModeling Complex Ecological Systems - An Introductiongautamsethi8561No ratings yet DLL - Mathematics 6 - Q2 - W7Document4 pagesDLL - Mathematics 6 - Q2 - W7Bryan Rex CabalNo ratings yet A Design of Rotor Bar InclinationDocument4 pagesA Design of Rotor Bar InclinationAnonymous Gm1md6kbNo ratings yet Answer Key - Final Exam - Math102-L02-Fall 2020-V1+V2Document10 pagesAnswer Key - Final Exam - Math102-L02-Fall 2020-V1+V2alwafa.q6rNo ratings yet Basics of GeometryDocument19 pagesBasics of GeometrybalachandarNo ratings yet SAP Doc - Types of Tables in SAPDocument8 pagesSAP Doc - Types of Tables in SAPmari_scripcaruNo ratings yet Activity3 SaezJohnCarlo NW3CDocument2 pagesActivity3 SaezJohnCarlo NW3CHalcyon LizardoNo ratings yet